Friday 29 April 2016

SciFest 2016

Another year, another free lunch. I thought I was going to have to miss the annual gathering of young scientists at The Institute because of my jury summons.  I was not needed by the law on Tuesday but required to offer my services again on Friday; so Wednesday and Thursday were providentially free. Thursday was SciFest, for budding scientists, and I spent the morning talking to a succession of youngsters about their thrashing at the frontier of science. I've done this before 2015 - stats - 2014 - 2013 All mostly good fun: what is it about the legs of today's youth? Are they wasting away? If these youths are sitting in front of their poster waiting to be judged and an ou'feller with a JUDGE badge approaches, would it not be meet and appropriate to stand up? If you're not going to stand, then could you not at least not mmmble but enunciate c l e a r l y?  If I have to kneel down or bend double to hear what you have to say, while you sit at ease on a chair, then I'm pissed off before we've even started.

If you're in a room with 100 other [young] scientists, would you not have a quick scoot round to see if anyone else is doing something interesting? Science goes in fads and phases, like in the Student Enterprise awards when everyone is making Cupcakes one year and Christmas decorations the next. Two groups in this year's SciFest were measuring the effect of music on cognition and memory. Bob the Matchmaker forced one lot to shift their duff up the hall to meet the other group and they seemed to start chatting way. OTOH, two groups, separated by only 4m were investigating the effect of sugar on reaction time and concentration.  I though one lot had better kit (iPad with reaction time software) but the other had better ideas [although I didn't make that invidious comparison] and so it was a complementary marriage in heaven.  But neither group seemed the least bit interested in talking to the 'rival' company. Conclusion? Teenage boys are less socially adept than girls? more complacent?Although one of these boy groups were the only kids on the day who stood up at my approach and proffered a hand to shake.

Numerous claims were made to me which really didn't stand up to scrutiny because - as ever and always - there was no statistical analysis. Sample sizes in general were small and so you'd get some data like:

Year 1
Year 3
Row Tot
The conclusion / claim would be that 4/6 = 67% was bigger than 1/5 = 20%, and so the Year 1 cohort was happier than the Year 3s: which is mathematically true but statistically silly. You'd need a sample at least 3x bigger and in the same proportion to show, with convincing / reproducible truth, that the rate of happiness decreased as you get older.  Just lash the numbers into the Easy Chi-Square Calculator, lads. I told two groups of young WITS [women in tech & science] that they should take their ideas forward but if they were going for national Young Scientist competition they'd need a) a bigger sample [treating their work to date as a pilot study] b) to bone up on some statistics. Google up "Chi-squared", I said. Science teachers, including ourselves, have a fixed idea that statistics is hard and so we should spare student scientists from ChiSq, t-test, Anova until they are older.  I disagree and assert that biological science without stats is biological stamp-collecting at best. The mechanics of Chi-squared are so simple they can be done on a scrap of paper with a pencil and the test is a powerful item in the crap-detector's toolkit CDT.

As it happens two of the presentations at SciFest 2016 have produced data that is really worrying if true  . . . and someone should be told!  One chap has found 25% toluene in certain cosmetics. Toluene is streng verboten in nail-varnish by the EU, but allowed in the US. Seemingly some US manufactured products are dumped on the European market: if the EPA, or the appropriate consumer watch-dog, doesn't know about this then they are remiss in their brief.  It may be that this is Euro-weenie hysteria or, contrariwise, carelessness with neurological health of US women.  Either way, it's worrying that EU regulations are being flouted.  I think that chap won one of the day's prizes.

The other consumer watch-dogs had taken 1g of potato chips/crisps and popped it in a home-made bomb-calorimeter to calculate a radically different value for the calories in a packet of crisps from the number that appeared on the packet. I was over-excited and misheard their spiel as saying that there were 2x more calories in the packet than in the crisps. That's the urban legend about the nutritional value of the cardboard being higher that the cornflakes they contain. Another project in the room seemed to claim that Cheerios are 90g sugar /100g material but that cannot be true - it's only 22%. I suggested that they should collate their results into a neat table and send them off to Tayto with a polite note asking if they were missing something in the methodology because the results were so discrepant. I suggested that they might score a box full of Tayto crisps from the publicity department at the company.  It wasn't until later that I realised that crisp manufacturers [not just Tayto!] might be consciously underestimating the calories to make it seem that their product was all fun and no fat. So Pringles was more likely to send a hit-man to a school in Co Kildare than send a cases of their wholly artifical potato-shaped food product. Those two projects were altogether much more interesting than the more usual bunch of teenagers asking the opinion of the other kids in their class / school.  How do you feel about crisps? Who cares! show us your fat content or carcinogens.

Apart from ranting on about statistics, I made a point of insisting that each team put their names on their presentation poster to show that they took ownership of their data and were proud of what they had achieved. If the name was on at all, it was usually added as a sort of footnote - and for them I suggested that immediately under the title would be a better location for the authors.  I passed by a stall later and saw that three names had been added to the top of the poster with a black marker.  It looked horrible, but it was nice to see that somebody had been listening to my suggestions.

No comments:

Post a Comment